OpenAI pauses reinforcement learning training; Frontier RL run on hold

OpenAI says it temporarily slows model scaling, including a two-week pause in reinforcement learning training on latest deploy-intended models, while its largest planned Frontier RL run remains on hold pending safety validation.